Pancreatic cystic lesions (PCL) are an increasing clinical challenge to clinicians and patients alike. Their prevalence is estimated to be around 16% in patients undergoing cross-sectional abdominal imaging for unrelated indications, discovered incidentally 1. While a large number of such incidental PCL may be small ( 15 mm and growth ≥ 2.5 mm/year. Amongst those patients with a baseline cyst size of 40 mm. While pancreatoduodenectomy is performed with low mortality and reasonable morbidity in most centres, it still remains a high-risk intervention. Indeed, risk of surgery may outweigh the survival benefit of resecting IPMNs, in particular when the resection is performed for LGD lesions that are unlikely to undergo malignant transformation. The study also showed that almost a quarter of patient undergoing pancreatic resection for LGD lesions developed severe postoperative complications (defined as Clavien-Dindo grade ≥ 3) and that the overall 1-year mortality was 2.6%. Furthermore, although the 5-years OS rate for LGD was 97%, this does not take into consideration the implications on patients' health related quality after major surgery 4. Taken from the above, we currently face two diagnostic conundrums in the management of small, incidental PCLs. Firstly, are these indeed IPMNs warranting surveillance? In the study by Levink et al., 21% of resection specimens were other cystic aetiologies, including serous cystic neoplasia (SCN) and pseudocysts—lesions that do not warrant surveillance (nor resection) in the first place. The observation of misclassification is supported by a recent systematic review 5 showing that one-in-every-four PCLs are preoperatively misclassified and, furthermore, that use of a preoperative EUS did not reduce the rate of cyst misclassification and, of note, this finding was consistent across health care systems around the world 4. Hence, the misclassification and low accuracy appear to be a consistent and a universal challenge. Secondly, in radiologically confirmed IPMNs, how well can we differentiate an IPMN with LGD from lesions with HGD or early invasive cancer? Although the risk of HGD/PC in small cysts with slow growth is miniscule, it is not non-existent and we are likely to miss patients who would benefit from early surgery although the numbers needed to treat are high. New strategies of risk evaluation, surveillance strategies and indication for surgery are urgently needed, likely with the use of multi-omics panels defining biological risk of the PCL. A way forward would be genomic analysis of cyst fluid exploring heterogeneity in known mutations to the IPMN lesions, such as GNAS and KRAS which are more prevalent in IPMNs with LGD with convergent evolution in mutations in RNF43 and TP53 leading to development of HGD and invasive cancers 6, a suggestion that is also supported by the revised Kyoto guidelines. This early diagnosis is imperative as firstly survival from resected invasive IPMN is considered better than pancreatic ductal adenocarcinoma 7, with 10-year survival rates approaching 30% and secondly, the role of adjuvant chemotherapy on survival benefit in invasive IPMN is not fully established 8 and a resection at earlier stages of the diseases undoubtedly confirms survival benefit 3. The challenge lies in the selection of which patients and lesions to consider for this invasive investigation (biopsies by EUS), the accuracy and representativeness of tissue accrual, and the capacity for mutation testing (usually by next-generation sequencing platforms) across the health care spectrum. Although the understanding of surveillance strategies and their outcomes for IPMNs is improving, considerable challenges lie ahead with an increasing volume of incidental and indeterminate PCLS being detected and entering surveillance programmes. The reliance on conventional modalities such as MRI and EUS has the challenge of interobserver variability at small cyst sizes and will meet capacity issues at the current incremental request for surveillance. It is likely future surveillance protocols will be determined by non-invasive modalities such as next-generation artificial intelligence and machine learning algorithms AI 9 to identify occult invasive foci or radiologically high-risk lesions that warrant surveillance. For situations in which a diagnostic conundrum still exists, it is possible that DNA/RNA based next generation sequencing panel of IPMN cyst fluid (or biopsies) will likely determine the need for surveillance 10. The authors declare no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Pandanaboyana et al. (Fri,) studied this question.